College Football Mathematical Preview: The rest of the Power Five

This month, I have been reporting on the results of my annual simulation of the upcoming college football season. As I explained in my overview of Michigan State’s schedule and in my breakdown of the Big Ten, I conducted a 100,000 cycle Monte Carlo simulation that uses published preseason rankings as an input.

The simulation allows me to generate odds for each game, which, among other things, results in projected standings, assuming that the favored teams always wins. This is the single “most likely” scenario (yet one of a truly astronomical number of scenarios). In reality, the underdog wins in about 25 percent of all college football games.

So, I also run a separate simulation that specifically looks for possible disruptions (usually in the form of key road upsets) and which generates a reasonable number of upsets. This “disruptive” scenario will at times give a different set of division or conference champions.

Now that we have a good handle on the Big Ten, it is time to focus on what the simulation has to say about the remaining Power Five conferences: the SEC, Big 12, ACC and the Pac-12. Let’s begin down south.

SEC Overview

Table 1 below summarizes the results of my simulation for the SEC. The table shows the consensus preseason ranking of each team, the total number of expected wins, projected records (based on both the “most likely” and “disruptive” scenarios), my strength of schedule calculations, and the odds for each team to win their division, the conference, make the playoffs, and win the national title.

As a reminder, my method to estimate strength of schedule involves a calculation of the number of expected wins that an average Power Five team would have with the schedule in question. Note that Auburn is very close to my benchmark of an “average Power Five team.” As such, Auburn’s expected win total (6.96) is almost identical to the strength of schedule (6.98).

In the SEC East, Georgia (ranked No. 5) is the highest ranked team based on the preseason magazines and therefore the team with the highest odds to win a division title (59 percent). Florida (No. 14) is second at 27 percent with no other team having odds above six percent.

A quick look at the preseason rankings and the strengths of schedules provides a clue as to why. The SEC East, outside of the top two teams, is predicted to be quite weak. If these preseason rankings are to be believed, the third best team in the SEC East (Kentucky, No. 44) would be the worst team in the SEC West.

Based on the projected point spreads, Georgia is likely to be favored in 11 of the 12 games on the schedule. The Bulldogs’ only loss in this scenario would be in the opening weekend neutral site game against Clemson (No. 2), as they draw Auburn (24) and Arkansas (40) as SEC West crossovers. In this most likely scenario, Georgia would easily win the East.

Florida, is not so lucky. The Gators draw Alabama (No. 1) at home and LSU (16) on the road and project to be underdogs in the both game. As always, the Florida/Georgia contest is played at the semi-neutral site of Jacksonville, Florida. The most likely outcome is for Florida to finish at 9-3.

However, in my disruptive simulation, Florida upsets both Georgia and Alabama and wins the division with a 7-1 record, including a win over Georgia. In this scenario, the Bulldogs drop a road game at Auburn and finish 6-2 in SEC play, a game behind the Gators.

In the SEC West, Alabama is the consensus preseason No. 1 team, and as a result, the Crimson Tide have the best odds to win the division (37 percent) with a most likely record of 12-0, edging out Texas A&M (30 percent and 11-1) and LSU (12 percent and 10-2). However, the disruptive simulation paints a slightly different picture.

The key factor is that Alabama has a much tougher schedule than Texas A&M (No. 7) and must travel to College Station to face the Aggies this year. That single game might be the most impactful game on the entire college football schedule. As noted above, Bama also has road games at Florida (No. 14) and at Auburn (No. 24) in addition to a neutral site game in Atlanta against Miami (No. 15) to open the season.

Meanwhile, Texas A&M’s toughest non-conference game is versus Colorado (No. 60). The Aggies do travel to LSU (16) and Ole Miss (26), but their SEC crossover games are against Missouri (45) and South Carolina (79).

Now, my simulation often predicts chaos in the SEC because there are so many good teams. Thus, I hesitate to pick against Alabama and instead pick Texas A&M to beat Florida in the SEC Title Game. That said, the schedule does suggest that this scenario has a very real possibility. The most likely conference championship matchup is still Alabama beating Georgia, but the SEC race looks to be the most interesting of the Power Five.

Big 12 Overview

Table 2 below gives an overview of the simulation results for the Big 12, using the same format as above.

The Big 12 does not have divisions. Instead, the teams play a full round robin of nine conference games, culminating in a conference title game between the top two teams. Making the championship game in the Big 12 is akin to “winning the division.”

In this case, the analysis is pretty simple. The projected top two teams in the conference, Oklahoma (No. 3) and Iowa State (No. 6), have the best overall odds to square off in the Big 12 Conference Championship game. Oklahoma is projected as the narrow favorite to win this contest.

There appears to be greater than a 50 percent chance of this particular matchup. Not only are the Sooners and Cyclones predicted to both be very good, both teams also grade out with the easiest two schedules in the conference.

Part of this is due to the fact that that Oklahoma and Iowa State do not have to play themselves. But beyond that, Oklahoma only has one true road game against a team projected to be in the top half of the league: Oklahoma State (No. 21). In the disruptive simulation, the Cowboys win this game, but the Sooner still easily earn a spot in the Big 12 Championship.

As for the Cyclones, they do travel to Norman to face Oklahoma this year (which is their only loss in both simulations), but their next most difficult road game appears to be at West Virginia (No. 31). So, while teams like Texas (19), Oklahoma State or TCU (25) certainly have a shot to rise up and sneak into the Big 12 Championship game, the conference schedule is working against them this year.

Pac-12 Overview

Table 3 below gives an overview of the simulation results for the Pac-12.

In the Pac-12 North the projected best team in the conference, Oregon (No. 9), has a clear advantage in overall odds to both win the division (57 percent) and the conference (34 percent). In the most likely scenario, Oregon runs the conference table and only loses to Ohio State in the non-conference. However, the disruptive simulation likes Washington’s odds.

The Huskies (No. 22) draw the Ducks in Seattle this year, and otherwise have a very manageable schedule. Washington’s only other projected top-35 conference opponents (UCLA, No. 33 and Arizona State, No. 20) must also travel to Seattle. In the disruptive simulation, the Huskies are the team to run the table in conference play, with their only loss of the season coming in Ann Arbor to the Michigan Wolverines in September.

The Pac-12 South appears to be more competitive, with a total of four teams ranked between No. 15 and No. 35 in the preseason: USC (17), Arizona State (20), Utah (27), and UCLA (33). The Trojans have a slight edge in odds to win the division (39 percent), just ahead of the Sun Devils (29 percent), the Utes (19 percent) and the Bruins (nine percent).

The Trojans are blessed with the easiest conference schedule in 2021, as USC draws Utah and UCLA at home this year and Arizona State on the road. USC also somehow avoids both Oregon and Washington as crossovers and thus are projected to finish 8-1 in the conference with a sole loss to Arizona State in both simulations.

In the disruptive simulation, Utah also goes 8-1, thanks to an upset win at home versus the Sun Devils, but the Utes would still lose the tiebreaker due to a road loss to USC. Arizona State, with road games at UCLA, Washington and Utah, will have a tough time keeping up in the standings. Similarly, UCLA must travel to Utah, USC and Washington.

Based on the disruptive simulation, I would project USC to beat Washington in the Pac-12 Championship game, but with a 11-2 final record, including a loss at Notre Dame.

ACC Overview

Finally, Table 4 below gives an overview of the simulation results for the ACC.

In the ACC Atlantic Division, Clemson (No. 2) is projected to once again run away with the division title. The raw odds give the Tigers a 73 percent chance to claim the division and a 44 percent chance to win the conference.

The only other Atlantic Division team projected to be in the preseason top 40 is North Carolina State, and the Wolfpack happen to have drawn the second toughest conference schedule with both North Carolina (No. 9) and Miami (No. 15) on the docket from the Coastal Division. In contrast, Clemson will face Pittsburgh (42) and Georgia Tech (72) as crossovers. Even if NC State were to upset Clemson in Raleigh, the Tigers look to have a cushion of several games in the division race.

Based on the preseason rankings the Coastal Division appears to be more competitive. That said, North Carolina (9) has the best odds (46 percent) due in large part to owning the easiest conference schedule. The Tar Heals host their biggest challenger, Miami (15), while their toughest road games are at NC State (34) and Virginia Tech (35). In both simulations, UNC’s only loss comes at Notre Dame.

That said, Miami is projected to win all of its conference games outside of the road game to Chapel Hill. Overall, the Hurricanes have a 30 percent chance to win the division. The Oct. 16 contest between North Carolina and Miami has the look of a winner-take-all affair. In the final analysis, Clemson is likely to beat North Carolina in the ACC Championship game with no more than one loss.

Notable Power Five Games

Based on this analysis, the following games will most likely have the biggest impact on the division and conference races, as well as on the teams that are candidates for the playoffs.

• Sept. 4: Georgia vs. Clemson (-2)
• Sept. 4: Alabama vs. Miami (+8)
• Sept. 4: LSU at UCLA (-3.5)
• Sept. 11: Iowa at Iowa State (-10)
• Sept. 11: Oregon at Ohio State (-6)
• Sept. 11: Washington at Michigan (push)
• Sept. 18: Alabama at Florida (+4)
• Oct. 2: Arizona State at UCLA (+1.5)
• Oct. 9: Georgia at Auburn (+6)
• Oct. 9: Utah at USC (-8)
• Oct. 9: Alabama at Texas A&M (push)
• Oct. 16: Miami at North Carolina (-7)
• Oct. 16: Arizona State at Utah (push)
• Oct. 23: USC at Notre Dame (-7)
• Oct. 30: Florida vs. Georgia (-3.5)
• Oct. 30: North Carolina at Notre Dame (-2.5)
• Nov. 6: Oregon at Washington (+3.5)
• Nov. 6: USC at Arizona State (-2)
• Nov. 13: Texas A&M at Ole Miss (+6)
• Nov. 20: Iowa at Oklahoma (-5.5)
• Nov. 20: Oregon at Utah (+5.5)
• Nov. 27: Texas A&M at LSU (+1.5)
• Nov. 27: Oklahoma at Oklahoma State (+6)

If Clemson is the ACC Champion, but does lose to Georgia in the first week of the season, will that still be enough to make the playoffs? What about Alabama if it loses to Texas A&M and fails to make the SEC Championship game?

In the next installment of this preseason series, I will take a quick tour of the Group of Five conference races and then make some projections as to the teams that will make the College Football Playoff, and the other New Year’s Six games. Until then, enjoy, and Go Green.